7-Motion discrimination technique for forearms using real-time EMG signals

Abstract

The objective of this study is to develop a method of discriminating real-time motion from electromyogram (EMG) signals. We previously proposed a motion discrimination method. That could discriminate five motions (hand opening, hand closing, hand chucking, wrist extension, and wrist flexion) at a rate of above 90 percent from four channel EMG signals in real time (the discrimination processing time was less than 300 ms). This method prevented elbow motions from interfering with hand motion discrimination, but it could only discriminate five motions, as opposed to a method using neural networks that can discriminate more than six. Here, we propose a real-time 7-motion discrimination method using a hyper-sphere model. The seven motions are hand opening, hand closing, hand chucking, wrist extension, wrist flexion, ulnar flexion, and radial flexion. The proposed model can learn EMG signals in real-time. Experimental results showed that the discrimination accuracy of this method was above 90 percent. Moreover, elbow motions did not interfere with the hand motion discrimination. The discrimination processing time was less than 300 ms.